Aims/hypothesis: The Human Genome Project seeks to identify all genes with the ultimate goal of evaluation of relative expression levels in physiology and in disease states. The purpose of the current study was the identification of the most abundant transcripts in human pancreatic islets and their relative expression levels using Serial Analysis of Gene Expression.
Methods: By cutting cDNAs into small uniform fragments (tags) and concatemerizing them into larger clones, the identity and relative abundance of genes can be estimated for a cDNA library. Approximately 49,000 SAGE tags were obtained from three human libraries: (i) ficoll gradient-purified islets (ii) islets further individually isolated by hand-picking, and (iii) pancreatic exocrine tissue.
Results: The relative abundance of each of the genes identified was approximated by the frequency of the tags. Gene ontology functions showed that all three libraries contained transcripts mostly encoding secreted factors. Comparison of the two islet libraries showed various degrees of contamination from the surrounding exocrine tissue (11 vs 25%). After removal of exocrine transcripts, the relative abundance of 2180 islet transcripts was determined. In addition to the most common genes (e.g. insulin, transthyretin, glucagon), a number of other abundant genes with ill-defined functions such as proSAAS or secretagogin, were also observed.
Conclusion/interpretation: This information could serve as a resource for gene discovery, for comparison of transcript abundance between tissues, and for monitoring gene expression in the study of beta-cell dysfunction of diabetes. Since the chromosomal location of the identified genes is known, this SAGE expression data can be used in setting priorities for candidate genes that map to linkage peaks in families affected with diabetes.